Determination of Optimal Corrosion Logging Frequencies for Gas Storage Fields

Abstract

Abstract Many of the existing wells in UNGSFs were originally designed to extract hydrocarbons, using construction techniques and practices that differ from modern storage well drilling and completion practices employed today. In addition to the age of these wells, their primary flow strings and tubulars may have been exposed to naturally occurring corrosive environments that could have degraded their wall thickness and remaining burst strength. This presents a serious risk of well failure and loss of containment for both the wells and reservoirs of these fields. To address this issue, the Pipelines and Hazardous Materials Safety Administration (PHMSA) in its interim final rule (IFR), recommended that all operators of gas storage facilities determine the reassessment intervals required to maintain the pressure rating and flow isolation properties of the flow string of all gas storage wells. This work presents a novel methodology for establishing the optimal logging frequencies required to maintain the integrity casing wall thickness of the primary flow string which serves as the leading indicator of casing integrity. In this study, metal loss data from magnetic flux-leakage logs run on existing wells gas storage wells are collected, vetted, and used to derive a probabilistic distribution of linear corrosion rates. Statistical values of corrosion rate corresponding to severity levels are used to approximate the corrosion rate distribution allowing for deterministic prediction of remaining burst strength with suitable failure criterion models. The Modified ASME B31G model is used to predict casing remaining life as a function of metal loss class and average corrosion rate. The reliability criterion for establishing fitness-for-service of a corroded casing string is evaluated separately using Monte Carlo simulation while reassessment intervals corresponding to a particular casing condition are calculated deterministically. The modeling results with the new approach presents predictions of remaining life for four commonly employed casing strings as a function of metal loss class and average corrosion rate. The results show that the remaining life of a corroded casing can be reasonably estimated as a function of static corrosion rate and metal loss class. The most significant finding is that a brand-new well with an average annual corrosion rate of 1.0 % initiating instantaneously upon installation will have an estimated remaining life of 28 to 32 years. Reassessment intervals are defined based on casing half-life as a function of metal loss class and average corrosion rate. A field case study is also presented to demonstrate the successful application of the new methodology in predicting reassessment frequency to mitigate risks from external corrosion mechanisms. The novelty of the proposed approach is in the ability to reasonably predict casing remaining life and safety target levels using derived statistical distributions of linear corrosion rates calculated from magnetic flux leakage logs. This in turn allows operators to confidently establish reassessment intervals for future casing inspection logging to timely identify corrosion threats in a manner that achieves a balance of risk mitigation and cost savings.